Authenticity is an important food quality criterion. Rapid methods for confirming authenticity or detecting adulteration
are increasingly demanded by food processors and consumers. Near infrared (NIR) spectroscopy has been used to detect
economic adulteration in pork . Pork samples were adulterated with liver and chicken in 10% increments. Prediction and
quantitative analysis were done using raw data and pretreatment spectra. The optimal prediction result was achieved by
partial least aquares(PLS) regression with standard normal variate(SNV) pretreatment for pork adulterated with liver
samples, and the correlation coefficient(R value), the root mean square error of calibration(RMSEC) and the root mean
square error of prediction (RMSEP) were 0.97706, 0.0673 and 0.0732, respectively. The best model for pork meat
adulterated with chicken samples was obtained by PLS with the raw spectra, and the correlation coefficient(R value),
RMSEP and RMSEC were 0.98614, 0.0525, and 0.122, respectively. The result shows that NIR technology can be
successfully used to detect adulteration in pork meat adulterated with liver and chicken.
In this study, the application potential of computer vision in on-line determination of CIE L*a*b* and content of
intramuscular fat (IMF) of pork was evaluated. Images of pork chop from 211 pig carcasses were captured while samples
were on a conveyor belt at the speed of 0.25 m•s-1 to simulate the on-line environment. CIE L*a*b* and IMF content
were measured with colorimeter and chemical extractor as reference. The KSW algorithm combined with region
selection was employed in eliminating the surrounding fat of longissimus dorsi muscle (MLD). RGB values of the pork
were counted and five methods were applied for transforming RGB values to CIE L*a*b* values. The region growing
algorithm with multiple seed points was applied to mask out the IMF pixels within the intensity corrected images. The
performances of the proposed algorithms were verified by comparing the measured reference values and the quality
characteristics obtained by image processing. MLD region of six samples could not be identified using the KSW
algorithm. Intensity nonuniformity of pork surface in the image can be eliminated efficiently, and IMF region of three
corrected images failed to be extracted. Given considerable variety of color and complexity of the pork surface, CIE L*,
a* and b* color of MLD could be predicted with correlation coefficients of 0.84, 0.54 and 0.47 respectively, and IMF
content could be determined with a correlation coefficient more than 0.70. The study demonstrated that it is feasible to
evaluate CIE L*a*b* values and IMF content on-line using computer vision.
This paper is aimed at investigating the possibility of sorting rice seeds by rapid techniques. Machine vision and dielectric separation were involved to determine external and internal quality of rice seeds. A conceptual rapid seed sorter is proposed. Two varieties of rice seeds planted and harvested in different years were involved in the experiments. Using morphological and color features gave a highly acceptable classification of normal and defective seeds. Dielectric parameters can be used to classify rice seeds into high vigor and low vigor. Combination of appearance characteristics and dielectric properties provide comprehensive response of seed quality. A highly acceptable defects classification and vigor improvement were achieved when the principle prototype was implemented for all the samples to test the adaptability. The good adaptability of machine vision and dielectric separation indicate the potential to determine quality of rice seeds rapidly. This paper presents the significant elements of the conceptual prototype and emphasizes the important aspects of the image processing and dielectric separation techniques.
The objective of this research is to develop a digital image analysis algorithm for detection of multiple rice seeds images.
The rice seeds used for this study involved a hybrid rice seed variety. Images of multiple rice seeds were acquired with a
machine vision system for quality inspection of bulk rice seeds, which is designed to inspect rice seeds on a rotating disk
with a CCD camera. Combining morphological operations and parallel processing gave improvements in accuracy, and a
reduction in computation time. Using image features selected based on classification ability; a highly acceptable defects
classification was achieved when the algorithm was implemented for all the samples to test the adaptability.
The objective of this research is to develop an intelligent information system to recognize rice seeds variety based on image processing and artificial neural network (ANN). At first, images of rice seeds were acquired with a color machine vision system. Each image was processed to extract twenty-one quantitative features. The classification ability of all the features was evaluated for different varieties recognition. To ensure system reasoning veracity and intelligence, ANN was used. A digital image-processing algorithm was developed to classify varieties of rice seeds based on external features. As an example of application of the system, the image data of common rice seeds varieties in Zhejiang province were input and the results of practical application proved that this software system achieved desired results.
A machine vision system for quality inspection of bulk rice seeds has been developed in this research. This system is designed to inspect rice seeds on a rotating disk with a CCD camera. The seeds scattering and positioning device on this system, under continuous feeding condition, reaches 85% fill-ratio of the number of holes on the disk. Combining morphological and color characteristics gave a highly acceptable classification. The high classification accuracies obtained using a small number of features indicate the potential of the algorithm for on-line inspection of germinated rice seeds in industrial environment. The overall average classification accuracy among the four categories was above 90%. This paper presents the significant elements of the computer vision system and emphasizes the important aspects of the image processing technique.
The objective of this research is to develop algorithm to recognize clipped stigma traces in rice seeds using image processing. At first, the micro-configuration of clipped stigma traces was observed with electronic scanning microscope. Then images of rice seeds were acquired with a color machine vision system. A digital image-processing algorithm based on morphological operations and Hough transform was developed to inspect the occurrence of clipped stigma traces. Five varieties of Jinyou402, Shanyou10, Zhongyou207, Jiayou and you3207 were evaluated. The algorithm was implemented with all image sets using a Matlab 6.5 procedure. The results showed that the algorithm achieved an average accuracy of 96%. The algorithm was proved to be insensitive to the different rice seed varieties.
The objective of this research is to develop algorithms to classify varieties of rice seeds based on external features. The rice seeds used for this study involved five varieties of Jinyou402, Shanyou10, Zhongyou207, JiayouandIIyou3207. Images of rice seeds were acquired with a color machine vision system. Each image was processed to extract twenty-two quantitative features. The classification ability of all the features was evaluated for different varieties recognition. The shape difference between Jinyou402 and Shanyou10 is obvious. The classification of Jinyou402 and Shanyou10 achieved an accuracy of 100% when a single feature such as the length-width ratio was used. Jinyou402 and IIyou couldn't be classified very well using one or two features. Then a perceptron was created and achieved an accuracy of 100% for both of Jinyou402 and IIyou. The shape difference between Jinyou402 and Zhongyou207 is obscure with naked eyes. All features were analyzed with principal components analysis method. A two-layer back propagation network was created and trained using gradient descent with momentum and adaptive learning rate. Nr. of hidden nodes was tested and early stopping skill was used. The total error of the finally established net is 2% for the classification of Jinyou402 and Zhongyou207. At last, all the images of five varieties were recognized as five classes. Another feed-forward network was created and trained using conjugate gradient back-propagation with Polak-Ribiere updates. Samples were disordered to train the network. The network achieved an average accuracy of about 85% for the five varieties.
The objective of this research is to develop a digital image analysis algorithm for detection of diseased rice seeds based on color features. The rice seeds used for this study involved five varieties of Jinyou402, Shanyou10, Zhongyou207, Jiayou99 and IIyou3207. Images of rice seeds were acquired with a color machine vision system. Each original RGB image was converted to HSV color space and preprocessed to show, as hue in the seed region while the pixels value of background was zero. The hue values were scaled so that they varied from 0.0 to 1.0. Then six color features were extracted and evaluated for their contributions to seed classification. Determined using Blocks method, the mean hue value shows the strongest classification ability. Parzen windowing function method was used to estimate probability density distribution and a threshold of mean hue was drawn to classify normal seeds and diseased seeds. The average accuracy of test data set is 95% for Jinyou402. Then the feature of hue histogram was extracted for diseased seeds and partitioned into two clusters of spot diseased seeds and severe diseased seeds. Desired results were achieved when the two cancroids locations were used to discriminate the disease degree. Combined with the two features of mean hue and histogram, all seeds could be classified as normal seeds, spot diseased seeds and severe diseased seeds. Finally, the algorithm was implemented for all the five varieties to test the adaptability.
This study was undertaken to develop computer vision-based rice seeds inspection technology for quality control. Color image classification using a discriminant analysis algorithm identifying germinated rice seed was successfully implemented. The hybrid rice seed cultivars involved were Jinyou402, Shanyou10, Zhongyou207 and Jiayou99. Sixteen morphological features and six color features were extracted from sample images belong to training sets. The color feature of 'Huebmean' shows the strongest classification ability among all the features. Computed as the area of seed region divided by area of the smallest convex polygon that can contain the seed region, the feature of 'Solidity' is prior to the other morphological features in germinated seeds recognition. Combined with the two features of 'Huebmean' and 'Solidity', discriminant analysis was used to classify normal rice seeds and seeds germinated on panicle. Results show that the algorithm achieved an overall average accuracy of 98.4% for both of normal seeds and germinated seeds in all cultivars. The combination of 'Huebmean' and 'Solidity' was proved to be a good indicator for germinated seeds. The simple discriminant algorithm using just two features shows high accuracy and good adaptability.
Obtaining clear images advantaged of improving the classification accuracy involves many factors, light source, lens
extender and background were discussed in this paper. The analysis of rice seed reflectance curves showed that the
wavelength of light source for discrimination of the diseased seeds from normal rice seeds in the monochromic image
recognition mode was about 815nm for jinyou402 and shanyou10. To determine optimizing conditions for acquiring
digital images of rice seed using a computer vision system, an adjustable color machine vision system was developed.
The machine vision system with 20mm to 25mm lens extender produce close-up images which made it easy to object
recognition of characteristics in hybrid rice seeds. White background was proved to be better than black background for
inspecting rice seeds infected by disease and using the algorithms based on shape. Experimental results indicated good
classification for most of the characteristics with the machine vision system. The same algorithm yielded better results in
optimizing condition for quality inspection of rice seed. Specifically, the image processing can correct for details such as
fine fissure with the machine vision system.
Incompletely closed glumes, germ on panicle and disease are three characteristics of hybrid rice seed, which are
actual reasons of poor seed quality. To find how many and which categories should be classified to meet the demand of
produce actually, the effects of various degree of incompletely closed glumes, germ on panicle and disease on ratio of
germination in changed storage periods were studied with standard germination rate test. An electronic scanning
microscope was used for micro-observation and measurement. Then the possibility of automation inspection was tested
with a machine vision system. The measures of increasing quality of hybrid rice seed were discussed in the paper at last.
In the light of the periods of treatment and the classification of characteristics, difference steps should be taken. Before
storage, Seeds with germ or severe disease should be rejected at first. Then seeds with incompletely closed glumes or
spot disease might be stored separately for a shorter time in dried condition and treated with antisepsis before using for
some special fields with lower quality demand. The seeds with fine fissure between glumes should be stored in a strictly
controlled condition separately and inspected before use, just like other normal and healthy seeds.
Incompletely closed glumes, germ and disease are three characteristics of hybrid rice seed. Image-processing
algorithms developed to detect these seed characteristics were presented in this paper. The rice seed used for this study
involved five varieties of Jinyou402, Shanyou10, Zhongyou207, Jiayou and IIyou. The algorithms were implemented
with a 5*600 images set, a 4*400 images set and the other 5*600 images set respectively. The image sets included black
background images, white background images and both sides images of rice seed. Results show that the algorithm for
inspecting seeds with incompletely closed glumes based on Radon Transform achieved an accuracy of 96% for normal
seeds, 92% for seeds with fine fissure and 87% for seeds with unclosed glumes, the algorithm for inspecting germinated
seeds on panicle based on PCA and ANN achieved n average accuracy of 98% for normal seeds, 88% for germinated
seeds on panicle and the algorithm for inspecting diseased seeds based on color features achieved an accuracy of 92% for
normal and healthy seeds, 95% for spot diseased seeds and 83% for severe diseased seeds.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.